Making static and interactive hexagon plots

Hexagons can be seen as a mixture of scatterplots and heat maps. Instead of points like we would have in a scatterplot, we get exclusively hexagons. Instead of color gradients standing for a third variable like it would be in a heat map, colors tells how many points approximately each hexagon bears. Let me highlight that this is a huge simplification of the hexagon plots.

Also, think about the hexagon plots as feasible alternative to over plotting. Through this recipe, we will learn how to brew hexagon plots using ggplot2. In the end, it will coerce this static form to an interactive one by using the plotly package. To make the point about over plotting, current Recipe is using air pollution data (robustbase::NOxEmissions). We are plotting the log of hourly mean NOx ambient concentration (ppb) against the square root of wind speed (meters/second).

Getting ready

Data about pollution comes from the robustbase package. As some packages require, we must...